The Alpine Schists and the Upper Triassic of Harpers Pass (Sheet S52), South Island, New Zealand

Transactions and Proceedings of the Royal Society of New Zealand, Volume 80, 1952, Page 213

The Alpine Schists and the Upper Triassic of Harpers Pass (Sheet S52), South Island, New Zealand By H. W. Wellman,* G. W. Grindley,* and F. W. Munden * Geological Survey, Greymouth. [Read before the Wellington Branch, October 11, 1951; received by the Editor, October 16, 1951] Contents Part I. Upper Triassic Beds at Trent River, North Westland Introduction Topography Narrative Description of the Fossils Part II. Relation of Upper Triassic Beds to the Alpine Schists The Zone of Decreasing Metamorphism The Metamorphic Belts Upper Sub-schist Belt Middle Sub-schist Belt Lower Sub-schist Belt Non-foliated Chlorite Schist Belt Foliated Chlorite Schist Belt Structure of the Metamorphic Belts Part III. The Lower Mesozoic Geosyncline Comparison of Rates of Sedimentation Relation between the Monotis Belts and the Lower Mesozoic Geosyncline Depth of burial of Monotis Belts Acknowledgments References Summary Upper Triassic (Noric) rocks with the index fossil Monotis richmondiana are described from near the alpine divide at Trent River, North Westland. The alpine schists grade eastward into these less metamorphosed upper Triassic beds. By the use of field criteria this zone of decreasing metamorphism has been divided into five metamorphic belts. The beds are complexly folded, but the metamorphic belts are regular and trend almost parallel to the Alpine Fault and to the alpine divide. These alpine upper Triassic beds are thicker than the better-known and abundantly fossiliferous beds of the same age at Nelson, Southland, and west Auckland. It is suggested that the alpine belt was deposited near the centre, and the abundantly fossiliferous beds near the western margin of a major geosyncline. The marginal belt is less metamorphosed than the alpine belt, and this difference in metamorphism may be due to the marginal belt having been less deeply buried than the alpine. Part I—Discovery of Monotis at Trent River, North Westland Introduction In 1946, one of the writers (F. W. M.) found a single cast of the fossil Monotis richmondiana in the Trent River area, North Westland (Sheet S52). Monotis is widely distributed in New Zealand, particularly in the South Island, and, being an index fossil confined to beds of Noric age, it has considerable importance as a guide to the pre-Tertiary structure. In the South Island, Monotis has been found in three widely separated areas—Nelson, Southland, and Canterbury

(fig. 1). In Nelson and Southland the Monotis-bearing beds are part of a fairly regular and moderately fossiliferous Permian-Triassic sequence, and Monotis has been reported from numerous localities. In Canterbury, Monotis has been found at a few places in the “greywackes” of the Southern Alps and eastern foothills. These “alpine greywackes” continue over the alpine divide into Westland, where with gradually increasing metamorphism they pass westward into the alpine schists. Monotis had not previously been recorded from Westland nor had it been recorded so close to the alpine schist. For these reasons the discovery of Monotis at Trent River has considerable geological importance. Topography From Mt. Aspiring to the Spenser Mountains the alpine divide is a well-defined and continuous high ridge. The Southern Alps proper are generally considered to end at Harpers Pass, a few of the mountains between the pass and the Spenser Mountains rising above 6,000 feet. The area described (fig. 2) includes part of this less-elevated region, and extends from Harpers Pass north to the Ahaura River. It forms part of the Sheet S52 area, and is partly in Westland and partly in Canterbury Land District. The northern part, which is entirely within Westland, is drained by Haupiri and Trent rivers. Haupiri River flows north-west almost directly away from the alpine divide. Trent River takes a less direct course. It rises close to a low saddle near the head of Haupiri River, and flows east towards the alpine divide for four miles. It then swings north-east and flows parallel to the divide for eight miles to join the Tukaekuri River six miles above its junction with the Ahaura River. The headwaters of the Trent and the Haupiri provide a convenient section from the alpine divide to the Alpine Fault at the west margin of the Alps. The southern boundary of the area is a major fault—the Taramakau-Hope Fault—which extends east beyond the area being described to Hope River. The Taramakau and Hurunui rivers follow this fault, the Taramakau flowing west-south-west to the west coast, and the Hurunui in the opposite direction to Lake Sumner. The low saddle of Harpers Pass marks the point where the fault crosses the alpine divide. Glaciers widened and deeply eroded the major river valleys. Post-glacial rivers have filled the deeper parts of the valleys with gravels, and the glacial features are largely obliterated even where the rocks are resistant to erosion. Along the Taramakau-Hope Fault the rocks are intensely shattered and post-glacial erosion has destroyed all evidence of glacial erosion. Most of the mountain peaks are between 5,000 and 6,000 feet in height. The upper timber line is at about 3,500 feet, and these peaks and the ridges that connect them are covered by grass and scrub. With the exception of the Ahaura, lower Trent, Taramakau, and Hurunui valleys, the lower part of the area is covered by forest. Most of the tracks follow close to the major rivers, near the middle of the glacial valleys and at some distance from the bedrock in the valley sides. In these rivers outcrops are rare. The smaller streams flow in gorges, and outcrops are abundant. Narrative In November, 1949, we made a five-day trip to Trent River to make a more detailed examination of the upper Triassic beds and to find out their relation to the alpine schists. The area that was examined in most detail is shown by

figure 4. On the first day we motored to the end of the Ahaura Valley road, then walked to our first camp at the junction of the Trent and Tutaekuri rivers. On the second day we continued up Trent River, established camp seven miles upstream at the mouth of Graf Creek, then made a circuit from Graf Creek over the top of Mount Monotis, and returned to camp by a steep gulch (Monotis Gulch) that joints Trent River a quarter of a mile upstream from Graf Creek. During this circuit two further Monotis specimens were found near the original locality. On the third day the alpine divide to the east was examined. We left the valley floor at an alluvial cone a mile upstream from the mouth of Graf Creek, followed the small stream (Confirmation Rill) that forms the cone, and then climbed through the bush to the open tops of the main divide. Two further Monotis specimens were found in loose but almost certainly locally derived boulders about 1,000 feet above Trent River. The main divide was followed south almost to Harper Pass at the heads of Taramakau and Hurunui rivers, and photo panoramas taken at three places. These panoramas were used in constructing the detail map (fig. 4). About a mile south of Confirmation Rill a large scree extends from the divide to within a thousand feet of the river and provides an easy route down from the tops. Thick volcanics and thin limestones are well exposed in the upper part of the stream that drains this scree. On the fourth day the excellent section exposed in the gorge of Trent River was examined. From the upper part of the Trent valley we crossed the Trent Saddle and camped in the east branch of Haupiri River about two miles upstream from the junction of the two branches. On the fifth and last day we walked down the Haupiri Valley to the Haupiri Road and returned by car to Greymouth. Two of us (H. W. W. and F. W. M.) had earlier made short trips to other parts of Sheet S52, and later two of us (G. W. G. and F. W. M.) checked the offset of rock boundaries shown by McKay (1893) along the Taramakau Valley. Description of the Fossils The main purpose of our trip was to search for the upper Triassic fossil Monotis; but, as mentioned, only four more definite specimens were found: two from the original locality (G.S. 4986) and two from Confirmation Rill (G.S. 4985), bringing the total to five. Four are faint impressions of single valves. In the solid rock they were merely partings only slightly weaker than the joint or bedding planes. The fifth and largest has two valves with a cavity between. The mudstone matrix is extremely compacted and apparently fresh, but no trace of calcite now exists. The first fossil was found on the face of a solid rock ledge. With the exception of a worn fossil from Confirmation Rill, the others were found by laboriously breaking the rocks along bedding planes as for graptolites. The secret of success is to find the correct rock facies (dark, well bedded but not fissile, siltstone or mudstone) and break the rock along its bedding planes. One fossil per man-hour of breaking is good going. The fossils are illustrated by fig. 10. Part II—Relation of the Upper Triassic Beds to the Alpine Schists The Zone of Decreasing Metamorphism Geologists have long known that a belt of schist extends south-west along the western side of the Southern Alps. Followed south-west, the belt swings south and connects with the wide Otago schist arch. Followed north-east, it dies out near Lake Rotoroa, but reappears about fifty miles north-east in Marlborough on the north side of the Wairau Fault. The alpine schist belt is clearly shown

on the 1947 geological map of the South Island (published 1949), its western boundary being the Alpine Fault (see Wellman and Willett, 1942). Neither the age of the rocks that form the schists nor the age of the metamorphism is known. The general structure of this alpine schist belt is fairly regular. High rank schists close to the Alpine Fault progressively decrease in metamorphism eastward away from it. The schist iso-grads parallel the equally regular strike of the schistosity planes and almost parallel the line of the fault. Metamorphism continues to decrease east beyond the schist belt through a belt of sub-schists and then stays much the same for many miles to the east. The line marking the limit of this progressive decrease in metamorphism lies close to the alpine divide from Mount Cook north to the Spenser Mountains. This paper is concerned with a part of this zone of decreasing metamorphism which has not previously been geologically examined. McKay (1893) was the first to map and describe the zone in a description of that part of the Alps lying between the Taramakau and Hokitika rivers. In his map the Alpine Fault is represented by the almost straight contact between the granites to the west and the schists to the east. He mentioned the topographic depression along this fault line, but did not map it as a fault. McKay mapped the following divisions in the zone of decreasing metamorphism. East Triassic Carboniferous Devonian (semi-metamorphic) Upper Schists Magnesian rocks or Mineral Belt Middle and Lower Schists West McKay based his ages on metamorphic and lithologic correlation with fossiliferous rocks elsewhere in New Zealand, and not on direct fossil evidence. The northern end of McKay's map adjoins the area described here and his Triassic beds form the southern extension of the Monotis-bearing beds of Trent River. Therefore his age for this group is correct. On McKay's map all the formations west of the Triassic boundary are offset along the line of the Taramakau River. This offset was vertified and measured. It is about five miles near the Alpine Fault, possibly decreasing to the east as indicated by McKay. Bell and Fraser (1906) mapped and described the zone of decreasing metamorphism from the Taramakau River south for twenty miles to Browning Pass. The next thirty miles of the belt south to Wanganui River was mapped by Morgan (1908). In both these descriptions the rocks within the zone of decreasing rank are mapped as the Arahura Series, with a line dividing the less metamorphosed part from the schists. In 1927 Henderson and Fyfe mapped the zone of decreasing metamorphism in the Murchison district. The southern end of their area is thirty miles north of the area here described. The maps, but not the Bulletin covering this area, have been published. In their maps they divided the zone into three units: schists, semi-schists, and less metamorphosed rocks. They named the latter “Mt. Robert Series,” and considered it to be Triassic. In the following table our metamorphic zones are correlated with those of previous workers. Correlation with north-west Otago is based on the hand

specimen criteria described by Turner and Hutton. Correlation with West Coast localities is based on the comparison of hand-specimens, supplemented to the south by the single metamorphic boundary that has already been mapped. Table Giving Correlation of Alpine Formations in Terms of Degree of Metamorphism Morgan, 1908. N.Z.G.S. Bull. 6 Bell & Fraser, 1906. N.Z.G.S. Bull. 1 McKay, 1893, pp. 171–3 This Report Henderson & Fyfe (Maps) Turner & Hutton (for N.W. Otago) Trans R.S.N.Z, vol. 65, pp. 329 and 405 Greywacke and argillite Triassic I. Upper subschists (upper Triassic) Mount Robert Series (Triassic) Less metamorphosed part of Arahura Series Upper part Arahura Series Carboniferous II. Middle subschists (slaty cleavage) Chlorite 1 Sub-zone III. Lower subschist Glenroy Series upper part Devonian IV. Non-foliated mica schists Glenroy Series lower part Chlorite 2 Sub-zone Upper Schists V. Foliated mica schists Chlorite 3 Sub-zone Mica schists Arahura Series Lower part Arahura Series Middle Schists Chlorite 4 Sub-zone Lower Schists Gneissic and dark schists Arahura Series Gneissic Schists The Metamorphic Belts To show the relation of the Trent River area to the schist belt and the Alpine Fault, a sketch map (fig. 2) is presented on a scale of 4.3 miles to an inch of the whole of Sheet S52 (Harpers Pass). The Alpine Fault crosses the north west corner of this sheet. Although the geology of the whole sheet is shown, only the schists and Triassic rocks on the south-east side of the fault are described. These rocks are divided into five belts numbered I to V in order of increasing metamorphism. It is assumed that these belts increase in age with increasing metamorphism, but only in the less metamorphosed belts can this be proved by internal evidence. Metamorphism progressively decreases away from the Alpine Fault; no metamorphic jumps are known and metamorphic belts can be mapped only by accepting rocks of specific ranks as marking their limits. Rocks of these specific ranks can be recognised at different places without much difficulty if care is taken to compare rocks that were originally similar. It is less easy to describe these rocks so that other geologists can recognize them. Hutton and Turner (1936) have set up standards for a limited range of low-rank schists, and other petrologists had earlier set up standards based on index minerals for the higher-rank metamorphic rocks with which this paper is not concerned. But none have been proposed for the sub-schistoze rocks which occupy the bulk of the area described. For this reason we regard our boundaries as tentative and subject to modification if and when such standards are set up.

On the left-hand side of the following table is given the lithology of each of the belts before metamorphism, on the right-hand side the existing New Zealand nomenclature which combines lithology with degree of metamorphism: Lithology before Metamorphism Lithology after Metamorphism I. Upper Sub-schists (upper Triassic) Muddy sandstone, fine bedded siltstone and mudstone, calcareous sandstone, massive mudstone, tuffaceous mudstone, basic volcanics, limestone, tuffaceous limestone, and conglomerate. Greywacke, argillite, conglomerate, limestone and volcanics. Quartz veins common in greywacke, rare in argillite. Slaty cleavage absent. II. Middle Sub-schist (age uncertain) Muddy sandstone, mudstone conglomerate, chert (radiolarian?), tuffaceous mudstone. Massive greywacke, slate, phyllite. Quartz veins common throughout. III. Lower Subschist (age uncertain) Massive sandstone, siltstone, mudstone, conglomerate. Contorted “flinty” greywacke with many ramnifying quartz veins, and streaks and bands of silicified argillite, conglomerate. Slaty cleavage absent. IV. Non-foliated Schist (age uncertain) Sandstone with thin siltstone bands, no conglomerate or volcanics seen. Non-foliated schist. V. Foliated Schist (age uncertain) Sandstone and siltstone. Foliated schist. I. Upper Sub-schist Belt Distribution. The rocks south-east of a line about three miles north-west of the main divide are mapped as Upper Sub-schist. This belt extends north and south for many miles roughly parallel to the main divide and east for an unknown distance beyond it. It was examined in detail only in the Trent Gorge and the adjoining area (fig. 4). Content. With the possible exception of the limestone and volcanics, all the beds are typical geosynclinal sediments. They were rapidly eroded from an elevated area, quickly transported without much weathering, and then rapidly deposited. Again, with the exception of the limestone and volcanics, they show no mappable large-scale changes in lithology, the same limited range of lithologies being repeated again and again in a short distance. Irrespective of metamorphism, they differ considerably from the most rapidly deposited Tertiary sediments known in New Zealand. In the bulk they consist of medium to coarse sandstone mostly in three to six foot bands with interbedded thinner bands of mudstone and siltstone. More conspicuous are the conglomerate and grit bands. These are common in the river gravels, but were found in place only at the lower end of the Trent River Gorge, where they are at least 300 feet thick. The component pebbles—indurated sandstone and mudstone and quartz (or quartzite)—range up to half an inch through. The fossiliferous beds have the greatest interest. Near Mt. Monotis they consist of well-bedded dark siltstones which break slightly more readily along than across the bedding planes. In Confirmation Rill the loose fossiliferous blocks are very similar to those near Mt. Monotis, but perhaps slightly more massive. These rocks show no trace of slaty cleavage, though equally as fine grained as those in the next metamorphic belt to the west. When exposed, they slowly break down into conchoidal fragments, breaking along fractures which are largely independent of joint or bedding planes. Bell and Fraser (1906, p. 45) described a rock east of Browning Pass that may be from the southern continuation of this belt as “an exceedingly hard black mudstone, which, devoid of

lamination and breaking with a splintery fracture, often resembles a compact fine-grained basaltic rock.” Quartz veins are absent or extremely rare in the mudstones and siltstones, but are not uncommon in the sandstones. Volcanics were found in place on the east side of Monotis Gulch and in the stream that drains the large slip on the west side of Trent River. At Monotis Gulch the volcanics comprise 50 feet of hard, light green rock—probably an altered tuff. A boulder of tuff at this place contained a limestone band, a few inches thick. The following thicker section crops out in the stream below the slip on the opposite side of Trent Valley. It may represent the same band on the opposite side of a syncline. Feet Well-bedded sandstone and siltstone (greywacke and argillite) 1,000 Basic lava, tuff and tuffaceous sandstone 500 Limestone slightly tuffaceous 50 Gritty tuffaceous limestone 10 Tuffaceous sandstone 200 Dark mudstone with many quartz veins 100 Tuffaceous sandstone 300 Sandstone and siltstone 500 McKay (1881) discovered abundant Monotis at Okuku River (north Canterbury) in limestone associated with similar volcanics, but no definite fossils were found in the limestone at Trent River. Three miles north of Confirmation Rill on the main divide one of us (F. W. M.) found similar volcanics, and in the upper Taramakau Valley within three miles of Harpers Pass, McKay (1893, p. 135) found “cherts, diorite sandstone, diabase ash beds, often highly calcareous and frequently associated with red jasperoidal slates.” Neither of these localities was re-examined. In New Zealand, limestone is closely associated with volcanics in sequences of many different ages. The association is too direct to be accidental, but the controlling mechanism is uncertain. Fossil Content. The Monotis shells, which are confined to this belt, have already been described. Other possible fossils are flattened tubes of sandstone found in mudstone boulders in the lower part of the Trent Gorge. These tubes may represent worm casts. The tubes are gently curved, up to three inches in length, and elliptical in cross-section, the larger diameter—along the bedding planes—being a quarter of an inch and the lesser an eighth of an inch. With these tubes and apparently closely related to them, are fine radiating marks partly defined by pyrite. Their origin is uncertain. These fossils are distinct from the well-known annellid Torlessia mackayi. Thickness. As the structure is complex, the thickness of beds in this belt is uncertain. According to the cross-section (fig. 3), which gives the simplest interpretation of the field evidence, the total thickness within the mapped area is 8,000 feet. The cross-section indicates that the stratigraphic interval between the Monotis locality at Mt. Monotis and that at Confirmation Rill is about 5,000 feet, but the proved thickness of Monotis-bearing beds is much less—only 500 feet—the stratigraphic distance between the first and second collections at Mt. Monotis. II. Middle Sub-schists Distribution. A two-mile-wide belt of rock extending from Taramakau River north-east across upper Trent River, along the north-west side of lower Trent

River to Tutaekuri River is mapped as middle sub-schist. The approximate width of the belt is defined in the upper Trent River section. Content. The rocks are more metamorphosed than the lithologically similar upper sub-schists. Sandstone, siltstone, and mudstone form the bulk. Conglomerate boulders are not uncommon in upper Trent River and have probably been derived from this belt. Volcanics and associated sediments were seen only near Trent Saddle. Red slates crop out at the saddle itself and red jaspillite on the south side of the stream that flows into Haupiri River about half a mile west of the saddle. Green tuffaceous mudstone and sandstone occur nearby. Slaty cleavage, the most significant mappable feature of this belt, is conspicuous in all the mudstone and siltstone bands Cleavage is absent, not only from the less metamorphosed upper sub-schists to the east, but also from the more metamorphosed lower sub-schists to the west. The Trent Gorge section showed that the zone of transition from the uncleaved to the well-cleaved mudstone is narrow enough to be used as a map boundary. The incoming of cleavage cannot be accounted for by any change in the original nature of the rocks because the two belts differ little in composition. Nor is it likely to be due to increase in the degree of folding, because both belts have been strongly folded. It is suggested that the depth of burial of the rocks at the time of folding controlled cleavage. On this interpretation the uncleaved upper sub-schists were not buried deep enough to have been cleaved. Fossil Content. The fossil Torlessia mackayi has been reported to the south (Morgan, 1908, p. 80; Cox, 1877, p. 78) in rocks of about this rank. In spite of careful search the fossil was not found. Traces of a larger fossil are common in cleaved dark mudstone in locally derived screes at one place on the south side of the Trent Gorge south-east of Mt. Monotis. Evidently at one time they consisted of masses of calcite up to three inches in length and up to a fifth of an inch thick. Most of the calcite has been weathered out and the rock is calcareous only in patches. Although quite unidentifiable, the close similarity of all the fragments, and the presence of calcite in otherwise non-calcareous rocks is evidence for an organic origin. Many of the weathered surfaces show striations parallel to, and caused by the cleavage. Where these striations cross the layers of calcite the layer is given a prismatic appearance that could be confused with the prismatic structure of a Maitaia shell. The general appearance of the fragments is such as would be expected if the distortion of the large brachiopods illustrated by Wilckens (1927, plate IX) from Mt. St. Mary were to be carried several stages further. Although unidentifiable, these fossils are characteristic and may have value for correlation in the sparsely fossiliferous rocks of the Alps. One of us (H. W. W.) has found similar fossil traces in rocks of about the same rank in screes at the road near the summit of Arthurs Pass, twenty-two miles south-west. Thickness. As the structure is uncertain, the total thickness of the beds can only be roughly estimated. The belt is about two miles wide. Graded bedding determined at many places in the Trent Gorge showed that about two-thirds of the outcrops have top to the east. Therefore, as the beds are all near vertical, their total thickness is probably not less than 3,000 feet, one third of the width of the belt. III. Lower Sub-schists Distribution. A three-mile-wide belt of rock extending from Taramakau River north-east across the headwaters of Haupiri River and through the Waikiti

Valley to Ahaura River is mapped as lower sub-schist. Rocks of this belt were examined at the mouth of Waikiti River, in the upper Haupiri River, and at Taramakau River. Content. The rocks consist of alternating bands of sandstone, siltstone, and mudstone, similar to those in the overlying belts. The increased rank of this belt is shown by an increase in silicification. The sandstone bands look massive and are extremely hard, the finer bands are more “flinty” than are corresponding beds in the upper belts. Contortion (“pinch” and “swell”) of the bedding is characteristic. Quartz veins are abundant in the sandstone bands in the upper part, and in both the sandstone and mudstone bands in the lower part. None of the beds are cleaved. A photo of a boulder from the lower part of this belt is illustrated by fig. 9. The rock consists of one to three inch bands of sandstone and darker siltstone. The beds are complexly deformed, minor puckers being superimposed on folds with a wave length of about six inches. Two sets of quartz veins are conspicuous—thin laminae that follow the bedding and thicker veins controlled by the fold axes. Other veins lie in different directions and show no obvious relation to the fabric of the rock. Many of the thin quartz veins are much more strongly sinuous than the bedding, and if these veins were considered alone they would give an erroneous idea of the degree of folding. Evidently the fabric of the rock controlled the formation of the quartz veins, the axes of the sinuosities being in the direction of the incipient linear schistosity. Large boulders of conglomerate were found in a small stream on the north side of Taramakau River, three miles upstream from the mouth of Otira River. They probably belong to about the middle of this belt, as they are associated with rocks of lower sub-schist rank in the stream. No conglomerate was found in the streams draining the belt to the north. None of the conglomerate pebbles have been stretched, but a few have been broken by pressure. The joints pass through matrix and pebbles alike. The matrix is welded to the pebbles to the same degree for each pebble. Most of the matrix can be separated without difficulty, but parts cannot be detached without breaking off the surface of the pebbles. No quartz veins were seen penetrating the conglomerate, but they are not uncommon in siltstone boulders in the same stream. The most striking feature is the apparent absence of pebbles of greywacke or of vein quartz, rock types equally as resistant as those represented and usually far more common. About a third of the pebbles are sedimentary, a third igneous, and the remainder too fine-grained for macroscopic determination. The igneous pebbles are pink, finegrained, somewhat gneissic, biotite granites ranging up to six inches and averaging one inch in diameter. They are better rounded and more spheroidal than the sedimentary pebbles. Two sedimentary types are represented—quartzites and siliceous limestones. Quartzite pebbles are the more numerous, and range from near white to medium grey in colour and up to three inches in diameter. About half are slightly calcareous and a few are veined with calcite. Siliceous limestone is less abundant and the fragments, on the average, smaller. The largest limestone pebble is flat, extremely siliceous, veined with calcite, and light brown-grey in colour. Smaller limestone fragments show up in the matrix with acid. No fossil traces were seen in either the limestone or the quartzite. The pebbles represent resistant rock types and cannot be considered representative of the rocks of the area of origin, of which they may represent only a minor part. In New Zealand similar quartzites, calcareous quartzites, and

siliceous limestones are known from the lower Palaeozoic—from the Devonian of Reefton and Baton River and from the Ordovician of North-West Nelson and Fiordland. These lower Palaeozoies may have formed the area of origin, but this is uncertain until the pebbles of the conglomerate yield fossils. IV. Non-foliated Mica-schists Distribution. A two-mile-wide belt of rock extending from the Taramakau River north-east past the head of Crooked River and through the middle part of Haupiri Valley to the Ahaura Valley downstream from the Waikiti River, is mapped as non-foliated schists. The belt was examined in the Taramakau, Haupiri, and Trent rivers. Content. This belt corresponds in rank with the Chlorite 2 sub-zone and the upper part of the Chlorite 3 sub-zone of Turner and Hutton (1936). All grades of sediment have been rendered schistose. Foliation is absent in the upper part and rare in the lower part. Compared with the less metamorphosed belts, bedding is poorly defined. However, the original nature of the beds—mostly thin-bedded sandstone and siltstone—is still obvious, and at favourable places it was possible to hazard a guess from graded bedding as to the “right side up” of the beds. V. Foliated Mica-schists Distribution. A belt of rock extending with decreasing width from Taramakau Valley north-east past Crooked River through Haupiri Valley to Ahaura Valley is mapped as foliated schist. The belt adjoins the non-foliated schists on the east and is cut off on the west by the Alpine Fault. The belt was not examined in detail and the mineralogical rank of its western edge is uncertain. Content. This belt consists of mica-schists with coarse foliation and perfect schistosity. In its upper part bedding is occasionally visible, but in its lower part the greater metamorphism has mostly obliterated it. Within this belt to the south McKay mapped sills of metamorphosed ultra basic rocks. He called them “Magnesian Rocks” or “Mineral Belt.” On his map they are shown as an extremely narrow belt offset about five miles at Taramakau River. Bell and Fraser and Morgan mapped the ultrabasics from Whitcombe River to Taramakau River. They called them the “Pounamu Series.” On their maps they showed them not as a continuous belt but as isolated lenses elongated along the strike of the beds. On the south side of the Taramakau these lenses are confined to a milewide belt which roughly corresponds in position with the narrower belt mapped by McKay. Within the mapped area Wellman (1944, p. 234b) has reported ultrabasic sills on the north side of the Taramakau Valley (Orangapuku River and Soapstone Creek). As on the south side of the river these sills strike north-east parallel to the schist. The ultra-basic sills do not line up across the Taramakau River, those on the north side being offset five miles east with respect to those on the south side. The schists that enclose the ultra-basic sills are of the same rank on both sides of the river, and this offset is almost certainly due to horizontal displacement that has taken place on the Taramakau-Hope Fault. Ultra-basics have not been found in place immediately north of Taramakau River, but fragments are not uncommon in streams up to and just beyond Crooked River. No ultra-basic outcrops or fragments were seen between Evans River and Ahaura River. To the north, fragments are common in Nancy and Tass rivers. The most northerly ultra-basic outcrop in the alpine schist is at Mill Creek, a tributary of Maruia River (Henderson and Fyfe, Geological map of

Fig. 1—Locality Map showing location of Fig. 2 and relation between Monotis localities and Undermass Structure.

Fig. 2—Geological Sketch Map of Sheet S.32 (Harper Pass).

Fig. 3—Cross Sction A-A' and Map Reference.

Fig. 4—Map of upper Trent River showing structure and topography.

Fig. 5—Trent River from Main Divide. Mount Monotis at extieme left of photo. Fossil locality on ridge beneath. Fig. 6—Upper part Trent Valley from Mt. Monotis, showing glacial valley and post-glacial gorge. Trent Saddle is at the upper end of the grass flat at head of valley.

Fig. 7— Valley from 1.000 ft. above Harper Pass, Lake Suinner is Just hidden by right hand side of valley. The depression on the line of the valley beyond Lake Summer marks the Taramakau-Hope Fault. Fig. 8—Taramakau Valley from 1,000 ft. above Haiper Pass. Taiamakau-Hope Fault piobably follows left hand side of valley to eross spin at the bend as indicated by dashed white line. The Taramakau Fault follows down the valley.

Fig. 9—Boulder of lower sub-schist at mouth of Warkiti River Fig. 10—Monotis fossils from Mt. Monotis and Confirmation Rill. Plasticine mould of first fossil found.

Rahu Survey District). The schist that encloses the ultra-basic rock appears to have about the same rank for the whole length of the ultra-basic belt. Structure of the Metamorphic Belts In rocks as complexly folded and with as few fossils and indicator beds as those of the Southern Alps it becomes almost impossible to trace and map individual beds. The most that can be done is to find the most simple structural interpretation not inconsistent with the sum total of the observations and then to draw cross-sections based on this interpretation. Such a procedure is admittedly arbitrary, but advance will only be made by expressing the available information in as simple a manner as possible, and this is best done by means of a cross-section. In attempting to interpret the structure, we have accepted the following points as being well established: (1) Bedding planes are flat or gently dipping over extremely small areas, the dip at most places being greater than 60°. (2) Many of the beds are overturned, and at most places the determination of overturning (by finding the top side of the beds) is more important structurally than the determination of exact dips. (3) The structure is essentially similar in all sections normal to the strike. (4) Adjoining fold axes are essentially parallel. Minor folds are closely parallel to major folds. (5) Fold axes are parallel to cleavage planes. With the limited information available it was impossible to draw cross-sections unless some unproved assumptions were accepted. The following assumptions appear probable and were used: (1) The nature of the folding is not essentially different at a depth of a few miles from what it is at the ground surface. (2) The beds increase in age in the direction of increasing metamorphism. Faults are not uncommon, and are mostly steeply dipping. They may play an important part in the structure, but the throw could not be determined, and faults have had to be neglected in drawing the cross-section. At all places attempts were made to find the stratigraphic top of the beds. Most of the alpine sub-schists are rhythmically bedded, and graded bedding is the most useful criterion. In the middle sub-schists slaty cleavage was used also, and found to agree with graded bedding. Twenty observations of graded bedding were made, fourteen indicating top to the east and six indicating top to the west. About half of each group of observations are reliable. The remainder are less certain, only one or two mud-sand alternations being exposed. The “top-bottom” observations are reasonably well distributed throughout the upper and middle sub-schists and are probably roughly representative. We expected the beds with top to the east to differ in average dip from those with top to the west, but this difference proved to be somewhat less than expected. Those with top to the east ranged from 45° west through vertical to 60° east and averaged 87° west, the bulk of these beds being either overturned or vertical. Those with top to the west ranged from 60° west to 80° west and averaged 71°, none being overturned. The average dip of the fold axes should lie between these two averages and be about 75° west. The cleavage planes do not support this, but dip at a lower angle—from 45° west to 70° west. In the detail map of the Trent River area the “top-bottom” beds are distinguished from those in which the apparent dip

only was observed, and the fold axes shown on this map are based on “top-bottom” beds. They are the minimum number of possible fold axes rather than the true number, which is likely to be greater. Only one fold axis was observed directly. In the middle sub-schists at Trent Gorge the limbs do not progressively flatten towards the axes but meet sharply. If this is true generally, the wide range in dip of the beds cannot be due to flattening of the limbs towards fold axes. More probably, it is due to folding of the fold axes. The structure of the Alps is known to be similar for many miles to the north and south of the area examined, and it can be reasonably assumed that the structure exposed at the ground surface does not differ appreciably from that at a depth of a few miles. The beds in this area vary in dip and in direction of stratigraphic top, and similar variations are likely in all cross-sections at right angles to the strike. All the observations have been projected on to a single line of section from Bell Hill through the Trent River Gorge to a mile east of Harper Pass. From the western edge of the upper sub-schists to the conglomerates at the lower end of the Trent Gorge the dip is vertical and the strike regular. Top is to the east. Therefore this section is the west flank of a syncline, and the conglomerate is many thousands of feet stratigraphically higher than the Monotis beds at Mt. Monotis. At Confirmation Rill, on the east side of Trent River, top is to the west and the beds form the east flank of the syncline. But the dip, both at this place and at other places on the steep side of the Trent Valley below the main divide is east at 45°. This 45° dip is inconsistent with the range of dips adopted for the axial plane of the folds, and may be due to surface creep caused by the steep slope. If the synclinal axis is correctly placed, and if the sequence is not interrupted by important faults, the two Monotis localities cannot be at the same stratigraphic horizon. As the simplest interpretation of the cross-section, the volcanics at Monotis Gulch have been tentatively correlated with those at Confirmation Rill. The evidence for the position of the easternmost anticline is indirect. As far as is known, metamorphism does not appreciably increase or decrease east of Trent River. If this is correct, the beds are unlikely to increase or decrease appreciably in age, and the next anticline (or fault) cannot be far east of the Trent Syncline. The section through the middle sub-schists is based on the Trent Gorge section (fig. 4); extra complexity is probable, but a more simple interpretation unlikely. The section through the lower sub-schists and the schists is diagrammatic. In this area it is tempting to assume that the folds progressively become more intense and more closely spaced in the direction of increasing metamorphism, and that the fold axes continue west with the same dip and pass into schistosity planes. But this interpretation meets with difficulties when the schist belt as a whole is considered. At Trent River and at other places on the east side of the schist belt the fold axes in the sub-schists appear to dip steeply beneath the schist belt. On the west side of the schist belt in Nelson and North-West Otago the fold axes in the sub-schists also appear to dip steeply beneath the schist belt. If a cross-section sketch is made across the full width of the schist belt it will be seen that it is difficult to postulate gently dipping fold axes in the central part of the schist belt where the schistosity planes are gently dipping. The relation of the fold axes to the schistosity planes is an important problem. It could perhaps be solved if cleavage planes and fold axes could be traced through to the lower sub-schists, a distance in the Trent River area of only three miles. But this will

not be easy, definite slaty cleavage being absent from even the most fine-grained lower sub-schists. The western end of the cross-section showing the dip of the Alpine Fault and the sub-surface structure to the west is purely diagrammatic. Part III—The Lower Mesozoic Geosyncline Comparison of Rates of Sedimentation The fossil Monotis richmondiana occurs at several different areas in the South Island, but in the North Island is known only from West Auckland. This fossil has a restricted upper. Triassic time range and it is reasonably certain that the Monotis-bearing beds represent the same time span in all parts of New Zealand. Figure 1 shows all the known localities of this fossil, the schist belt, and the Alpine Fault. It is convenient to consider these fossil localities as being parts of three belts: a western belt, a central or alpine belt, and an eastern belt. The western belt is represented by the abundantly fossiliferous beds at Southland, Nelson, and West Auckland. At these places the upper Triassic beds form part of a fairly well known lower Mesozoic sequence, which at Southland and West Auckland extends up into the Jurassic. The structure is relatively simple, the Mesozoic beds being folded into a major syncline, the axis of which is shown on the map. At most places along this belt the Monotis beds are abundantly fossiliferous (Marwick, 1935, p. 300) and the shells form a large part of well-sorted medium sandstone shell beds. It seems likely that the shell beds were concentrated from less abundantly fossiliferous material, and that the average rate of subsidence was small. The thickness of the Monotis-bearing beds is different in different parts of the western belt. The minimum at Nelson is a hundred feet or so, the average about a thousand feet, and the maximum two thousand feet. The alpine belt is close to the eastern margin of the schist and has a known length of about 200 miles, Monotis having been reported from Macauley River at the south (Speight, 1921) and from Lake Rotoiti at the north. The Trent River locality is within this belt, and is the only area in which anything is known of the structure. As already mentioned, the structure is complex and the full thickness of the Monotis-bearing beds is not known. Although the proved thickness is only 500 feet, the full thickness is certainly more, and probably as much as 5,000 feet. Younger beds have been eroded from the Trent River area, and the original thickness may have been even greater. Subsidence and deposition were evidently rapid along this belt. Fossils have been collected from only a few places along the alpine belt. In part, this apparent scarcity of fossils is due to lack of detailed examination. Nevertheless, the rocks at many places are extremely well exposed in bare mountains and if fossils are abundant, they should not be hard to find. Gravels derived from this belt are widely distributed over the lowlands on both sides of the Alps. That no fragments of shell beds have been found in these gravels increases the probability that shell beds are either rare or absent in the Alps. The extreme rarity of Monotis at Trent River has already been mentioned, and the fossil is probably no more abundant at the other localities along the alpine belt. It has been mentioned that in the western belt Monotis is found mostly in sandstone. In the alpine belt it has been found only in mudstone. This difference may be due to the different rates at which mudstone and sandstone accumulated in the

western and in the alpine belt. In the western belt the fossiliferous sandstone bands probably accumulated slowly. In the alpine belt subsidence was probably much faster with fewer interruptions, so that the sediment was rarely resorted and concentrated by currents. Under such circumstances the lighter muds would have been deposited slower than the heavier sands, and a foot of mudstone may represent a much greater interval of time than a foot of sandstone. Most of the shells are single valves which may have been transported into the alpine area, perhaps along the surface of the sea. Other things being equal, the abundance of such fossils in any particular bed would depend on the rate at which that bed was deposited. This may explain why they have been found in the slowly deposited mudstone and not in the more rapidly deposited sandstone. The eastern Monotis belt is about forty miles east of the alpine belt. Monotis has been reported from Malvern Hills at the south end by Speight (1920, p. 106) and from Okuku River at the north end by McKay (1881, p. 100). It has recently been found by one of the authors (H. W. W.) at Lees Valley between these two localities. This belt has not been examined in detail and nothing is known of the structure or thickness of the Monotis zone. The rocks are similar to those of the alpine belt, and consist of a monotonous succession of extremely compacted sandstone and siltstone. Submarine volcanics similar to those at Trent River are not uncommon, and were used by McKay as evidence for correlation between the two areas. Associated with the volcanics at Malvern Hills and Okuku River are lenses of limestone with the fossil Monotis. At Lees Valley the fossil is contained in medium sandstone, slightly more compacted, but otherwise similar to the Monotis-bearing sandstone of the western belt. Relation between the Monotis Belts and the Lower Mesozoic Geosyncline It is generally accepted that the lower Mesozoic beds, of which the Monotis zone forms a small part, were deposited in a major geosyncline that extended far beyond New Zealand. The evidence given above makes it possible to consider the growth of this geosyncline during the upper part of the Triassic. As the Triassic beds in all areas are clearly of shallow water origin, the geosyncline was subsiding at approximately the same rate as the sediments were accumulating. Therefore, in a given area the thickness of sediment deposited during a given time interval can be used as a measure of the amount of subsidence. Monotis richmondiana had a limited time range which was probably about the same in all parts of the New Zealand area. Therefore the rate of subsidence of different areas in the upper Triassic can be judged from the thickness of the Monotis-bearing beds. These thicknesses have already been mentioned and shown to be greater in the alpine belt at Trent River than at any place on the western belt. Therefore the most rapidly subsiding (axial) part of the geosyncline was probably east of the western belt and possibly not far from the alpine belt. Depth of Burial of Monotis Belts Judging from the youngest beds that have been preserved, the lower Mesozoic geosyncline continued to grow until the end of the Jurassic. It was then compressed and the beds folded. Erosion followed, and in most places a great thickness of the upper beds was eroded. Cretaceous and Tertiary beds were then deposited on the eroded surface cut across the upturned edges of the older beds. These younger beds have since been elevated and eroded so that the contact with the older beds below them is exposed at many places. Although the degree of compaction of the base of these younger beds is different at different places, at no

place is it as great as that of the older beds beneath. Evidently the thickness of the old rocks that had been eroded was greater in all places than that of the younger beds that were deposited and in part eroded later. Consequently these younger beds can have had no influence on the degree of compaction of the older, and if compaction and the associated features of low rank metamorphism are due to depth of burial, it is possible to estimate the depth of burial of the upper Triassic beds from their degree of compaction. As deposition in the geosyncline appears to have come to an end at the end of the Jurassic, the degree of compaction of the upper Triassic beds must be due to the thickness of the Jurassic, of which in most areas no trace now remains. In the western belt the upper Triassic beds are only moderately compacted. They contain bituminous coal at Nelson and probably at Southland. They are not intersected by quartz veins. The degree of alteration along the alpine belt is considerably greater. Immediately underlying beds contain slates, and the upper Triassic beds are intersected by quartz veins. In the eastern belt the degree of compaction appears to be slightly less than in the alpine belt, but somewhat more than in the western belt. Fragments of coal from several localities have been analysed and shown to be of low volatile bituminous or of anthracite rank. The greater metamorphism of the alpine Triassic suggests that the alpine area was covered by a greater thickness of Jurassic beds than either the western or eastern areas, and hence that it continued to subside most rapidly during the Jurassic. Acknowledgments We are grateful to Professor W. N. Benson for the reference to the Macauley River Monotis locality described by Speight and mentioned in “Recent Advances”; to Professor R. S. Allan for information about the Mt. Temple locality and for permission to record his own discovery of poor Monotis in the bed of Rakaia River just above Rakaia Gorge bridge. References Bell, J. M., and Fraser, Colin, 1906. The Geology of the Hokitika Sheet. N.Z. Geol. Surv. Bull. 1. Cox, S. H., 1877. Report on Westland District. Rep Geol. Explorations, 9, 63–93. Henderson, John, and Fyfe, H. E. Maps of the Murchison Subdivision. Hutton, C. O., and Turner, F. J., 1936. Metamorphic Zones in North-West Otago Trans. R.S.N.Z., 65, 405–6. Marwick, John, 1935. Some New Genera of the Myalinidae and Pterirdae of New Zealand. Trans. R.S.N.Z., 65 (3), 295–303. McKay, Alexander, 1881 The Older Sedimentary Rocks of Ashley and Amuri Counties. Rep. Geol. Explorations 13. —, 1893. Geological Explorations of the Northern Part of Westland. Mines Reports, C–3. Morgan, P. G., 1908. Geology of the Mikonur Subdivision. N.Z. Geol. Surv Bull 6 Speight, Robert, 1920. Two New Fossil Localities in Maitai Rocks. N.Z. Journ. Sci. & Tech., 3 (2), 105–6. Speight, Robert, 1921 Notes on a Geological Excursion to Lake Tekapo. Trans. N. Z. Inst., 53, 37–46. Wellman, H. W., and Willett, R. W., 1942. The Geology of the West Coast from Abut Head to Milford Sound. Part I. Trans. R.S.N.Z., 71 (4), 282–306. Wellman, H. W, 1944. Talc in North-West Nelson and North Westland N.Z. Journ. Sci. & Tech., 24 (5B), 227B-235B. Wilckens, Otto, 1927 Contributions to the Palaeontology of the New Zealand Trias. N. Z. Geol. Surv. Pal. Bull. 12.